A Unified Framework to Elicit Structured Feedback for Interpretable Multi-Trait Essay Scoring
Computation and Language
Summary
The authors created a new system called HiFTS that scores essays by looking at multiple related traits together, instead of separately. Their method generates detailed, structured feedback that is connected to the scoring criteria before assigning trait and overall scores. They trained their system using guidance from a large language model and applied special optimization to balance scoring accuracy and feedback quality. They also made a new Chinese essay dataset, CFMS-34, to test their approach. Results show HiFTS works well at both scoring and providing clear, rubric-aligned feedback.
Authors
Shihang Yang, Sanwoo Lee, Ningning Zhao, Yunfang Wu
Abstract
Multi-trait Automated Essay Scoring (AES) requires rubric-grounded reasoning across interdependent traits, rather than isolated score prediction. Existing feedback-enhanced methods often decouple feedback from scoring or assess traits independently, weakening score--feedback consistency and rubric alignment. We propose HiFTS, a unified autoregressive framework that generates hierarchical CoT feedback before predicting trait-level and holistic scores. HiFTS distills rubric-grounded hierarchical CoT feedback from a teacher LLM and trains student models to jointly generate feedback and scores. HiFTS further applies Group Relative Policy Optimization with a composite reward balancing score agreement, calibration, feedback quality, and structural validity. At inference, a lightweight global prior provides holistic guidance to reduce drift during long-form reasoning. We also introduce CFMS-34, a Chinese multi-trait AES dataset with 951 essays annotated with holistic scores and 34 rubric-based traits. Experiments on CFMS-34 and ASAP++ show that HiFTS achieves strong holistic and trait-level scoring while producing coherent, rubric-aligned feedback.